Sampling Methods In Power System Reliability Evaluation Study

Power system reliability is one of the important rods for the plan, design, operation and maintenance of power system. There are two basic methods for the reliability evaluation of power system. These two methods are the analytical method and Monte Carlo simulation, respectively. The analytical method is of the higher precision, but its computational scale becomes large in the index rule with increasing system part numbers. The analytical method is, therefore, suitable for the reliability evaluation of small-scale power system with simple structure. Monte Carlo simulation can be favorably used to evaluate the reliability of large-scale power system because of its computational quantities being independent of the system scale. The shortcoming of Monte Carlo method is that it will take a great deal of the computation time when the high computation precision is required. Due to the requirement for increasing the accuracy of evaluation algorithm in the project practice, to improve sampling algorithm so as to enhance efficiency of the evaluation has become an important topic with certain theoretical significance and practical value. In this paper, the general principles and algorithm of Monte Carlo method used for the reliability evaluation of power system have been described in detail and the several methods applied usually to reliability evaluation of power system are presented. These methods are average-scattered sampling method, importance sampling method and adaptive algorithm, respectively, and their principles and algorithms have been also outlined systematically. The analyses on the characteristics and applicable computing circumstance of above algorithms have been made. Taking IEEE-RTS as an example, which is regarded as the standard testing system, the state of system has been simulated based on the above methods. The simulation program has been constructed using C code and the large number of the calculations has been performed. All the methods mentioned above were compared systematically, according to various calculated results. Also, the mixed method with the combination of the analytical method and Monte Carlo simulation are proposed for the reliability evaluation of generating system. The mixed method is a simple and is of strong transplant. From the standard testing system IEEE-RTS, the evaluated results of reliability show the efficiency of the mixed method for speeding up the convergence of the simulated indexes under the required precision. Further, a combination of average-scattered sampling technique with mixed method is given for reliability evaluation of generating system. By the evaluated results for IEEE-RTS system, it is clear that this new method is useable for reducing the sampling variance and accelerating the convergence of iteration calculation.In summary, to adopt the improved approach of sample can increases the simulation speed of Monte Carlo method in the reliability evaluation of power system. The methods given in this paper can be widely used for large-scale power system.